SCIENTIFIC-LINUX-USERS Archives

September 2004

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Perret Yannick <[log in to unmask]>
Reply To:
Date:
Wed, 22 Sep 2004 17:08:04 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (127 lines)
Thank you for your help.

I finally found the problem: the machine on which I did my
tests (a brand new one, just out of the box) has its SCSI controler
broken...
As it was a new machine I don't think about a material problem...
By the way, the module do not gives (in this case at least) any
usefull error message.

Sorry for the thread and thank you Stephan.

I'm now trying to build a 64+32 bit version of SL302 through
the kickstart, which seems to be possible.
I will keep you informed if I succed in that installation.

Regards,
--
Yannick Perret



Stephan Wiesand wrote:

>I installed with PXE + kickstart. The initrd I used has its modules in
>the directory 2.4.21-15.EL/ia32e.
>
>I made the kernel and initrd used for installation available in
>http://www-zeuthen.desy.de/~wiesand/sl/
>
>Maybe this helps.
>
>Cheers,
>       Stephan
>
>On Wed, 22 Sep 2004, Perret Yannick wrote:
>
>
>
>>Thank for your answer.
>>
>>How do you perform the installation ?
>>
>>In my case, I'm just using PXE, as for all of the other machines.
>>It works fine for i386 and for x86_64 without that disk, but
>>kickstart failed "no disk found", with error messages about
>>missing symbols for the module.
>>
>>With the i386 SL, I can check that it is the same disk:
>>uname -a
>>Linux ccwl0003 2.4.21-15.ELsmp #1 SMP Wed May 12 19:48:28 CDT 2004 i686
>>athlon i386 GNU/Linux
>>lspci | grep -i scsi
>>02:04.0 SCSI storage controller: LSI Logic / Symbios Logic 53c1030 PCI-X
>>Fusion-MPT Dual Ultra320 SCSI (rev 07)
>>
>>
>>Stephan Wiesand wrote:
>>
>>
>>
>>>Both kickstart installs and the system itself worked fine on this hardware
>>>at least up to kernel 2.4.21-15.0.2.ELsmp:
>>>
>>>[wiesand@heliade4]~% uname -rm
>>>2.4.21-15.0.2.ELsmp x86_64
>>>
>>>[wiesand@heliade4]~% /sbin/lspci|grep -i scsi
>>>02:04.0 SCSI storage controller: LSI Logic / Symbios Logic 53c1030 PCI-X
>>>Fusion-MPT Dual Ultra320 SCSI (rev 07)
>>>
>>>[wiesand@heliade4]~% /sbin/lsmod|grep mpt
>>>mptscsih               43344   5
>>>mptbase                45280   3 [mptscsih]
>>>scsi_mod              126404   3 [sg mptscsih sd_mod]
>>>
>>>This is an Opteron, not a Nocona system.
>>>
>>>Cheers,
>>>     Stephan
>>>
>>>
>>>On Wed, 22 Sep 2004, Perret Yannick wrote:
>>>
>>>
>>>
>>>
>>>
>>>>Some additionnal notes:
>>>>
>>>>I checked a machine which is under SL302-64 bits
>>>>(also a dual-opteron, but from an other constructor).
>>>>By chance, this machine is IDE, and so I encounted
>>>>no problem during installation.
>>>>
>>>>I checked  the /lib/modules and found the previously
>>>>evil modules mpt*.o, and the second one called
>>>>mptscsih.o also failed to load due to unresolved
>>>>symbols. The other one used (mptbase.o) loads
>>>>without problems but cannot initialize because
>>>>it uses the previous one.
>>>>
>>>>So the question is:
>>>>is it a compilation mistake if this module does
>>>>not work ? Or is it because x86_64 does not
>>>>support it ?
>>>>
>>>>I'm trying to answer to the question by
>>>>rebuilding the src.rpm for the kernel with a
>>>>modified .config.
>>>>
>>>>Help/clues would be appreaciated.
>>>>
>>>>Regards,
>>>>--
>>>>Yannick Perret
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>>
>
>
>

ATOM RSS1 RSS2